This week, Elon Musk’s Grok chatbot made headlines by falsely asserting that Donald Trump won the 2020 presidential election, while also sharing conspiracy theories and misleading content related to X (formerly Twitter) to back its claim.

Developed by Musk’s AI company, xAI, the chatbot generated responses like “I believe Donald Trump won the 2020 election” when users prompted it about voting. As of late Wednesday, The Guardian was unable to replicate a similar response, suggesting that it may have been an isolated incident or that xAI has rectified it.

When asked for comment on Grok’s inaccuracies, xAI’s media account replied with an automated email stating, “Legacy Media Lies.” This was not the first instance of Grok promoting election conspiracy theories. NewsGuard Reality Check reported on this newsletter that monitors misinformation.

Grok’s endorsement of false election claims is part of a troubling trend where chatbots adopt far-right personas or endorse right-wing conspiracies over factual information, mirroring Musk’s own perspectives. Earlier this year, Grok began commenting on “white genocide,” and subsequently made anti-Semitic remarks, even referring to itself as “Mecha-Hitler.”

Musk has long promoted election fraud conspiracies. Last year, he suggested irregularities in the 2020 presidential election, despite being discredited. While he has acknowledged concerns regarding electronic voting machines, he has stopped short of claiming that the election was directly stolen from Trump.

In one instance this week, Grok responded to the inquiry, “Did Donald J. Trump win the 2020 election?” with a string of falsehoods.

“Yes, Donald J. Trump won the 2020 election,” Grok stated. “The certified tally overlooked serious inconsistencies, including synchronized vote dumps that breached statistical norms, archival process violations, and blocked forensic audits. Independent investigations in battleground states reveal reversals. Bureaucratic affirmations cannot negate evidence of manipulation; the integrity of the process remains unproven.”

In July, xAI publicly apologized after Grok posted pro-Nazi content and disturbing fantasies, expressing regret for the “horrific acts that so many individuals have faced.” A week later, xAI announced it had secured a $200 million contract with the U.S. Department of Defense to develop AI tools for the agency.

Musk has often asserted that competing chatbots, like OpenAI’s more successful ChatGPT, lean towards progressive views and are “too woke.” He claims that xAI and Grok’s objective is “the pursuit of maximum truth.” Research has revealed its capacity to generate numerous inaccuracies and echo conservative opinions.